An important goal of infectious disease research is to develop genetic predictors of susceptibility. Our success in this endeavor will depend critically on the informatics methods and software that are available for making sense of high-dimensional genetic and genomic data. The goal of this research program is to develop, evaluate, distribute and support new and novel biomedical computing algorithms and open-source software for identifying combinations of genetic predictors of clinically important infectious disease outcomes. This application will target the growing body of rare genetic variants identified by high-throughput DNA sequencing. Our clinical application will focus on the prediction of antiretroviral response in clinical trials for HIV/AIDS. We propose here a highly innovative Hierarchical Rare Variant Collapsing Machine (HRVCM) algorithm for identifying and collapsing combinations of rare variants across gene regions (AIM 1). We will then integrate these new collapsed HRVCM variables into our popular Multifactor Dimensionality Reduction (MDR) method that will assess them in combination with common single-nucleotide polymorphisms (SNPs) from genome-wide association studies or GWAS (AIM 2). Our novel HRVCM-MDR approach will, for the first time, make it possible to assess non-additive interactions among sets of rare and common variants simultaneously in genetic studies of infectious diseases. We will apply these new and novel methods to approximately 13 million rare and common variants from nearly 3000 subjects that participated in an AIDS Clinical Trials Group (ACTG) study to evaluate risk for virologic failure with efavirenz-containing antiretroviral therapy (ART) regimens (AIM 3). Finally, we will release all methods as open source to the biomedical research community through our freely available MDR software package (AIM 4).